part 1] THROUGH THE ANDES OF PEHU AND BOLIVIA. 23 



and have been sheared into innumerable small prisms and rhombs, 

 groups of which still show optical continuity. 



Distortion and fracture of the plagioclase is very well marked. A 

 feature of the rock is the extensive development of magnetite, in the form 

 of minute grains scattered throughout the hornblende and biotite. 



The second or dioritic phase was marked by the intrusion of 

 melanocratic rocks of basic character. 



Although for the greater part these have undergone profound 

 modification, their original composition can be determined from 

 some of the less altered examples which may be classed with the 

 augite-diorites. In many cases, however, the pyroxene has been 

 almost completely converted into fibrous hornblende. The almost 

 constant association of these unhanded rocks with the gneissic 

 modification of the granodiorite has ahead}" been noted. They are 

 unaffected by the forces which produced such intense deformation 

 in the latter rock, and are therefore regarded as constituting a dis- 

 tinct phase of intrusion, their localization having been determined 

 by lines of weakness which were evidently developed in areas of 

 maximum deformation. Although the dynamic stresses never 

 again attained such a degree of magnitude, the abundant quartz- 

 epidote veins which mark lines of fracture in both the diorite and the 

 older rocks show that the district was still subjected to considerable 

 earth-movements. The mineralosrical changes which have affected 

 the diorite over wide areas, as exemplified by the uralitization of 

 the pyroxenes and the saussuritization of the felspars, cannot be 

 ascribed to any mere process of weathering, or even to contact- 

 metamorphism : for the distinctive effects of the latter, to be 

 described later, are clearly exhibited where the rock is penetrated 

 by the acid magma of the third or granitic phase. 



It would appear, therefore, that their origin must be attributed 

 to some other cause, and the most plausible suggestion is that they 

 are the effects of waning dynamic action which was no longer 

 sufficiently great to produce macroscopic structural change. This 

 theory is supported by microscopic investigation. 



The most distinctive feature of the altered rock is the fact that, 

 when viewed with ordinary light, each individual crystal, almost 

 without exception, is clearly defined by an opaque dust-like margin. 

 Under a high magnification this is seen to be composed of minute 

 granules, which appear to be chiefly epidote. The rock, then, has 

 been rendered completely permeable to solution from the large 

 epidote-lined fissm-es with which it abounds, down to the finest 

 capillary cracks : the bulk of the epidote being evidently derived 

 from the pyroxene during its conversion into hornblende. 



This pronounced permeability, which is not possessed to the same 

 extent by the coarser granodiorite (although the latter contains 

 much disseminated epidote), appears to be due to incipient granu- 

 lation of an even -textured rock of relatively fine grain, as a result 

 of the action of dynamic stresses. 



Further evidence of internal movement is also afforded by the 

 biotite. which occurs in the form of somewhat large plates that 



